CH 7 Resp

7 The respiratory system
The function of the respiratory system is to exchange O2 and CO2 with the blood.
To understand this system from a bioengineering viewpoint, we will first discuss
the gross anatomy of the lungs and their associated structures, and then discuss the
mechanics of breathing.
7.1 Gross anatomy
We divide the respiratory system into two subsystems: the conducting airways and
the associated structures.
7.1.1 The conducting airways and pulmonary vasculature
The conducting airways form a fantastically complex branching tree designed to

transport air efficiently into the alveoli, the smallest air-filled structures in the
lung where blood/gas exchange takes place. Air enters through the mouth or nose
then passes through (in order): the pharynx (the throat), the larynx (the voice
box), and the trachea (the large tube passing down the neck). The trachea splits to
form two bronchi (singular: bronchus), each of which feed air to one of the lungs
(Fig. 7.1, color plate).
Each bronchus splits to form bronchioles, which, in turn, split to form smaller
bronchioles, and so on (Fig. 7.2). After about 16 levels of branching, we reach
the terminal bronchioles, which are the smallest structures that have a purely air-
conducting function, that is, in which essentially no blood/gas exchange takes place
(Fig. 7.3). In adult lungs, the structures distal to the terminal bronchioles consist of
several generations of respiratory bronchioles, alveolar ducts and alveolar sacs,
which collectively are known as the acinus; this is where the gas exchange occurs
(Fig. 7.4). All of the conducting airways are lined by pulmonary epithelial cells,
which we will see play a number of important roles in ensuring the proper function
of the lungs.
283 7.1 Gross anatomy
Figure 7.2
A resin cast of the airways of a pair of human lungs shows the fantastically complex geometry of the branching airways,
beginning from the trachea (T) and moving down into the bronchial tree (B). Note the branching of the peripheral airways
in the higher-magnification view in the inset. In the left lung, the arteries (A) and veins (V) have been filled with a slightly
darker resin. Reprinted by permission of the publisher from The Pathway for Oxygen: Structure and Function in the
Mammalian Respiratory System by E. R. Weibel, p. 273, Cambridge, MA: Harvard University Press [2]. Copyright C
1984 by the President and Fellows of Harvard College.
The alveoli (singular: alveolus) form after birth, by a fascinating process in which
the walls of the smallest air-containing passages in the lungs become scalloped so
as to enlarge their surface area. The number of alveoli gradually increases with
time until about the age of six, after which the lungs are essentially fully formed
(Fig. 7.5). As can be imagined, it is not easy to know exactly how many alveoli
there are. Early estimates [6] suggested that there are about 300 million alveoli,
with radii ranging from 75 to 300 µm, and mean radius of approximately 150 µm.
More recent estimates [7] have suggested that the average number of alveoli is
480 million, with an average radius of about 100 µm. The total surface area of
z Name D (mm) A (cm2) Number
0 Trachea 18 2.54 1
1 Mainstem bronchi 12.2 2.33 2
2 Lobar bronchi 8.3 2.13 4
3 Segmental bronchi 5.6 2.00 8

4 Subsegmental bronchi 4.5 2.48 16
…
12 Bronchiole 0.95 28.8 4,096
13… Bronchiole 0.82 44.5 8,192
16 Terminal bronchiole 0.49 225.0 65,536
17 Respiratory bronchiole 1 0.40 300.0 131,072

20 Alveolar duct 1 0.34 1,743.0 1,048,576
21 Alveolar duct 2 0.31 2733.0 2,097,152
22 Alveolar duct 3 0.29 5070.0 4,194,304
23 Alveoli 0.25 7530.0 8,388,608
Figure 7.3
Schematic diagram of the organization of the airway tree, showing the different functional zones (“conducting zone,”
“transitional zone,” and “respiratory zone”) as a function of generation numbers, z. The diameter (D), cross-sectional
area (A) and number of airways are shown in the colums on the right. The number of airways is actually larger than that
listed here, since the tabulated values are based on a simple bifurcating model of the airways; the reality is more
complex. T refers to the terminal generation, T-1 to the generation immediately preceding the terminal generation, etc.
The dimensions are suitable for a lung that is inflated to three quarters of capacity. The figure at left is from Weibel [3],
supplemented at right by entries from [2]. Reproduced with permission of Lippincott Williams & Wilkins.
the alveolar walls has been quoted to be as large as 140 m2 [2], which is about 75
times the body’s external surface area.
In addition to the conducting airways for air transport, it is necessary to have a
system for delivering the blood to the alveoli so that mass transfer can take place.
The lungs are supplied by the pulmonary arteries and drained by the pulmonary
veins. The smallest units of the pulmonary vasculature are the pulmonary capil-
laries, which run as closely packed units inside the alveolar walls. The net effect
is that a thin “sheet” of blood flows within much of the alveolar wall (Fig. 7.4).
Interposed between the blood and the air is a thin layer of tissue, consisting of
capillary endothelial cells, basement membrane, and airway epithelial cells. The
total thickness of this tissue varies, with an arithmetic mean thickness of 2.22 µm
and a harmonic (geometric) mean thickness of 0.62 µm [2]. The combination of an
enormous alveolar surface area, a comparably large pulmonary capillary surface
285 7.1 Gross anatomy
A B
a
a a
Figure 7.4
Overview of lung microanatomy. (A) Scanning electron micrograph of lung tissue, showing how a peripheral bronchiloe
(BL) branches into terminal bronchioles (T), which, in turn, divide to produce respiratory bronchioles and alveolar ducts
(arrows). This gazelle lung was fixed in such a way as to prevent collapse of the alveoli and terminal segments of the
bronchial tree. (B) A plastic cast of the alveolar capillary network within a small group of alveoli (a), photographed after
the surrounding tissue was dissolved away. The two images are not taken from the same sample or even from the same
species; the white box in (A) approximately corresponds to the field of view in (B) and is simply included to give a sense
of relative scale. Scale bar in panel (B) is 50 µm. (A) is reproduced from Fawcett [4] and Gehr et al. Journal of
Morphology 168:5-15, Copyright C 1981, and is used with permission of Wiley-Liss, Inc., a subsidiary of John Wiley &
Sons, Inc. (B) is Copyright C 1990 from Electron Microscopy of the Lung by Schraufnagel [5] and is reproduced by
permission of Routledge/Taylor & Francis Group, LLC.
area, and the very thin tissue layer between blood and air make the lung a very
effective mass transfer device.
7.1.2 Associated structures
As well as the conducting airways, the pulmonary system includes a number of

other tissues (the associated structures) that together are responsible for the inspi-
ration and expiration of air. The lungs are contained within the thoracic cage, con-
sisting of the ribs, intercostal muscles, sternum (breastbone), spine, diaphragm,
and neck and shoulder muscles.1 These together form a deformable “container” for
the lungs. Attached to the exterior surface of the lung and to the interior surface of
the thoracic cage is the pleural membrane. The space enclosed by this membrane
is the pleural space, which is normally filled with the intrapleural fluid (Fig. 7.6).
Each lung has its own pleura.
1
The intercostal muscles are located between the ribs. The diaphragm is the large muscle at the base of the lungs,
familiar to anyone who has been “winded.”
106 Na
605
400 Dunnill (1962)
Weibel (1963)
Davies and Reid (1964)
Angus and Thurlbeck (1972)
300
200
100
4 8 12 18 24 AGE
Birth Months 2 3 6 12 18 Adult
Years
Figure 7.5
Increase in the number of alveoli with age in human lungs. There is a gradually increasing trend until about age six, after
which the number of alveoli is essentially constant. The symbols show the results from different studies. The original
data sources can be found in Weibel [2]. Reprinted by permission of the publisher from The Pathway for Oxygen:
Structure and Function in the Mammalian Respiratory System by Ewald R. Weibel, p. 227, Cambridge, MA: Harvard
University Press. Copyright C 1984 by the President and Fellows of Harvard College.
Figure 7.6
Schematic of the thoracic cage, pleural membrane, intrapleural fluid, and lungs. This is not to scale; the intrapleural fluid
actually forms a very thin layer. Note that there is no communication between the right and left intrapleural fluids. After
Vander et al. [1]. With kind permission of The McGraw-Hill Companies.
287 7.2 Biomechanics of breathing
Inspiratory
positions
Diaphragm
Figure 7.7
Chest wall and diaphragm motion during inspiration. The solid and dashed lines indicate expiratory and inspiratory
positions, respectively. From Vander et al. [1]. With kind permission of The McGraw-Hill Companies.
7.2 Biomechanics of breathing
An interesting aspect of the breathing process is the fact that there are no muscles
attached to the lungs to aid with inflation and deflation. Rather, the necessary
forces are transmitted to the lungs through the intrapleural fluid via the mechanisms
described below. This system allows a uniform inflation of the lungs while avoiding
the high stresses and deformations that would result if a muscle inserted into the
soft lung tissue.
Inspiration and expiration proceed by two very different mechanisms, as
described below.
Inspiration. In quiet inspiration, the diaphragm contracts and moves downward,

thereby increasing the volume of the thoracic cavity. In addition, the external
intercostal muscles and the scalene muscles (at the neck) contract to move the rib
cage up and out (Fig. 7.7). This also increases thoracic cavity volume, although
the contribution of the intercostal and scalene muscles is minimal during quiet
inspiration. Since the intrapleural fluid is essentially incompressible, an increase
in thoracic cage volume must be accompanied by an increase in lung volume. The
best analogy is that of a balloon inside a fluid-filled container of variable volume
(e.g., a piston and cylinder). As the piston moves back and forth, the balloon’s
volume shrinks and grows. In forced inspiration the above description still applies.
However, the contribution of the intercostal and scalene muscles becomes more
pronounced.
Expiration. Quiet expiration is a passive process that relies on the natural
tendency of the lung to collapse (see Section 7.3 for detailed discussion). The
diaphragm, the external intercostal muscles, and the scalene muscles relax, allow-
ing the lung and the thoracic cage to decrease in volume. Forced expiration is an
active process. In addition to the relaxations of quiet expiration, the internal inter-
costal muscles contract to pivot the ribs downward and in. Also, the abdominal
muscles contract, pulling down the rib cage, increasing abdominal pressure, and
thereby forcing the diaphragm upwards. These mechanisms together act to reduce
thoracic cavity volume, which allows a large quantity of air to be quickly expelled
from the lungs.
7.3 Lung elasticity and surface tension effects
It was previously stated that the natural tendency of the lung is to collapse. Before
examining why this is the case, we briefly look at what this fact implies about the
intrapleural pressure.
At rest, when the subject is neither inspiring nor expiring, the pressure within
the conducting airways is essentially atmospheric.2 To counterbalance the lung’s
tendency to collapse, it is clear that the lung must be “blown up”. In other words,
the pressure in the lung must exceed the surrounding pressure, which in this case
is the intrapleural pressure. Therefore, at rest, intrapleural pressure must be slightly
negative. This is confirmed by measurements, with typical intrapleural pressures
of −3 to −4 mmHg (gauge) at rest. During inspiration and expiration, intrapleural
pressures deviate from this resting value. For example, during rapid inspiration,
intrapleural pressure can fall to as low as −100 mmHg, while it can become slightly
positive during forced expiration.
If the pleural cavity is opened to the atmosphere (for example in a puncture
wound), then the pressure difference required to inflate the lung is lost. This causes
lung collapse, or pneumothorax.
We now turn to the question of why the lung tends to collapse. There are two
reasons:
r Lung tissue (or parenchyma) is naturally elastic. It contains a very high propor-
tion of elastin (see Section 2.4). Under normal conditions, the lung is expanded
2
During inspiration, alveolar pressure is slightly negative (about −1.5 mmHg) to draw air into the alveoli, and it is
slightly positive during expiration.
289 7.3 Lung elasticity and surface tension effects
60 Fy = 0
sx (Lagrangian stress Fx /Aox), Fy = 10 × 103 dynes
Fy = 20 × 103 dynes
dynes/cm2, × 103
40 Fy = 50 × 103 dynes
20
0
0.8 1.0 1.2 1.4 1.6 1.8
lx (Stretch ratio L x/L ox)
Figure 7.8
Stress–strain curve for lung parenchyma under biaxial loading conditions. For constant transverse loads (differing Fy
values in dynes), the stress is plotted as a function of stretch in the x direction. Note the significant increase in tangent
modulus at stretch ratios above approximately 1.5. From Vawter and Humphrey [8]. With kind permission of The
McGraw-Hill Companies.
(inflated) so the lung parenchyma is under tension. The elastic restoring force
from this tension acts to decrease lung volume. It is interesting to note that
lung parenchyma is a strain-stiffening material (i.e., its incremental Young’s
modulus increases with strain, as shown in Fig. 7.8). Such behavior is typical of
connective tissues and reflects the arrangement and orientation of stress-bearing
fibers within the tissue (see Section 9.10.3). At low stresses, the collagen and
elastin within the lung, which are the main stress-bearing materials, are ran-
domly oriented. At high stresses they align along the direction of stretching
(Fig. 7.9). Additionally, at high strain, the individual collagen and elastin fibers
stiffen, because of their molecular structure, which resembles that of a helical
spring (see Section 9.10.1). The net effect is that the lung becomes significantly
stiffer as it is inflated.
r Surface tension acts to collapse the lung. To see why this is the case, consider
a single alveolus filled with air. In order to maintain the viability of cells in
the alveolar wall, the inner surface of the alveolus must be coated with a thin
fluid layer. This creates an air/fluid interface within the alveolus, as shown in
Fig. 7.10. Whenever an interface between two immiscible phases forms,
energy is required to maintain that interface. The system attempts to minimize
Undeformed state: low strain and low incremental Young’s modulus
Tension Tension
Deformed state: large strain and large incremental Young’s modulus
Figure 7.9
Schematic diagram of collagen and elastin orientation in connective tissue. As fibers align in the direction of stretching,
the tissue stiffens.
Air, pressure p i
Liquid film, pressure po
Alveolar wall (elastic)
Intrapleural pressure ppleural
Figure 7.10
A single alveolus, showing pressures in surrounding tissue, in the air, and within the liquid film lining the alveolar wall
(not to scale). See text for definition of symbols.
this energy by reducing the interfacial area. This gives rise to an effective
force that acts along the interface, called the surface tension. It is characterized
by a surface tension coefficient σ, the value of which depends on the two
contacting phases. In the single alveolus, a reduction in interfacial area can
only be accomplished by collapsing the alveolus. To counteract this tendency,
the pressure of the alveolar air ( pi ) must exceed the pressure on the other side
of the interface, namely the liquid film pressure po (see Box 7.1).
291 7.3 Lung elasticity and surface tension effects
Box 7.1 Analogy of an air bubble in liquid

The situation in an alveolus is directly analogous to the case of an air bubble in
an infinite volume of liquid. If the interfacial (surface) tension coefficient for
the air/liquid interface is σ , then at equilibrium Laplace’s law states that, for a
spherical bubble of radius R, the difference between internal air pressure pi
and the pressure in the surrounding fluid, po , is given by
2σ
pi − po = (7.1)
R
Using Laplace’s law it is trivial to calculate the pressure difference pi − po

in a single alveolus. Using Equation (7.1) with numerical values R = 150 µm =
1.5 × 10−2 cm and σ = 72 dynes/cm (the latter value being appropriate for an
air/water interface) yields pi − po = 9600 dynes/cm2 , or 7.2 mmHg. At rest, when
pi is atmospheric, this result implies a liquid film pressure of −7.2 mmHg.
These data can be used to make an inference about the intrapleural pres-
sure, ppleural . Because of the elasticity of the lung parenchyma and alveolar wall
tissue, the liquid film pressure must be greater than pleural pressure. Hence,
we predict that ppleural < −7.2 mmHg. However, we have stated that the mea-
sured resting value of ppleural is −3 to −4 mmHg, which contradicts the above
calculation.
The resolution of this apparent contradiction is that we have used the wrong
value of σ in Equation (7.1). The alveoli are in part lined by type II epithelial
cells, which secrete a surfactant that mixes with the alveolar liquid film and greatly
decreases σ for the resulting air/liquid interface. This, in turn, increases ppleural ,
effectively making it easier to overcome surface tension and so to breathe. The
value of σ for saline plus surfactant in contact with air has been measured (Figs.
7.11 and 7.12). The data show that the volume of σ depends on the extension
of the air/liquid interface. Specifically, σ is small at low relative interfacial areas
and is large at high relative interfacial areas. This means that σ is lower at the
start of an inspiratory cycle than at the end. Since large values of σ imply that
more negative intrapleural pressures are required for inspiration, it follows that
it is easier to inspire near the beginning of an inspiratory cycle than at the end.
Essentially, surface tension acts in a non-linear fashion to stiffen the lung near the
end of inspiration.
A notable feature of the surface tension data is the hysteresis present: at a given
relative interfacial area, the surface tension is higher when the interfacial area is
increasing than when it is decreasing. This means that surface tension makes it
more difficult to breathe in than to breathe out.
Figure 7.11
Surface balance apparatus used to measure surface tension by Clements. The shallow tray is filled with saline, and the
fluid to be tested is layered on top of the saline. A thin platinum strip (0.001 inch thick) is then suspended in the fluid
from the arm of a very sensitive strain gauge-based force transducer. The force exerted by the fluid on the platinum strip
is measured, from which the surface tension can be computed. To change the surface area of the tested fluid, a barrier
can be moved along the trough. When this is done cyclically while measuring the surface tension continuously, any
hysteresis effects in the surface tension versus area behavior are readily detected. Modified from Clements [9] with
permission of the American Physiological Society.
The relative contributions of lung parenchymal elasticity and surface tension

can be seen in Fig. 7.13, in which lungs are “inflated” either with physiological
saline or air. A number of comments about these data are in order:
r When the lung is inflated with saline, there are no surface tension effects and
therefore the curve labeled “2” represents the effects of parenchymal elasticity
alone. The stiffening of lung parenchyma at high volumes (high strains) is
evident. Also note that some hysteresis is present, as a result of viscoelastic
effects in the parenchyma.
r The difference between the curves labeled “2” and “1” indicates the importance
of surface tension effects. Significantly greater pressure is required to inflate
the lung when surface tension effects are present. A great deal of hysteresis is
evident, as expected based on the hysteresis in the value of σ shown in Fig. 7.12.
Clinically, surfactant is very important. It makes breathing much easier (less

pressure difference needed to inflate the lung), so that the diaphragm and intercostal
muscles can do less work with every inspiration. Premature infants are often born
with reduced surfactant-producing abilities, since the lungs are one of the last
organs to mature in utero (Fig. 7.14). These infants have a great deal of difficulty
breathing (a condition known as respiratory distress syndrome) and are typically
given synthetic surfactant until their type II epithelial cells mature.
293 7.4 Mass transfer
Lung
Detergent extract Plasma Water
Relative area (%) 100
60
20
10 20 30 40 50 60 70 80
Surface tension (dynes/cm)
Figure 7.12
Surface tension versus interfacial relative area for water, plasma, detergent, and lung extract. The surface tensions of pure
water and detergent are 72 and 22 dynes/cm, respectively, and show no dependence on interfacial area. Plasma and lung
extract, however, have a surface tension that depends on interfacial area and also on whether the interfacial area is
increasing or decreasing (arrows). This effect, leading to hysteresis in the curves, is particularly prominent for lung
extract. Modified from Clements [9] with permission of the American Physiological Society.
7.4 Mass transfer
The lungs are remarkable mass transfer devices that permit blood/gas exchange. It
is of interest to examine the performance of the lungs as mass transfer units, and
towards this end we will consider two points of view: (i) a microscopic one, in
which we focus on gas transport in a single acinus; and (ii) a macroscopic one, in
which we focus on the entire lung.
When considering mass transfer at the level of the acinus, we should keep in
mind that O2 transport from air to blood consists of several steps: O2 in the acinar
air must move until it is adjacent to the thin tissue layer separating blood from
air, must cross this thin tissue layer, and must then move through the blood to
300
2 1 3 4
200
Volume (ml)
100
CAT
0 4 8 12 16 20
Pressure (cmH2O)
Figure 7.13
Pressure–volume relationship for inflation of an excised degassed cat lung with air or saline solution. Loop 1 is an initial
inflation/deflation with air; loop 2 is an inflation/deflation with saline; loop 3 is an inflation/deflation with air after saline;
and loop 4 is an air inflation/deflation after the lung was washed with detergent. Note that in all inflation–deflation cycles
(but particularly when the lung is inflated with air), the deflation path differs from the inflation path. The difference
between loop 1 and 2 results from the absence of surface tension effects in the saline-filled lung. After saline filling,
some airways are collapsed and a significant opening pressure is required to inflate the lung, as can be seen by the
initially horizontal part of the inflation paths for loops 3 and 4. Filling the lung with saline removes surfactant, as does
washing with detergent, producing higher deflation pressures for loops 3 and 4. Modified from Bachofen et al. [10] with
permission of the American Physiological Society.
bind to hemoglobin in red cells. In thinking about these processes, it is convenient

to consider the “air-side” and “blood-side” mass transfer as individual processes
coupled by transport across the blood–air tissue barrier. In a well-designed system,
we expect that the mass transport capacity of these individual processes will be
approximately matched, although we will see that it is not easy to define what
“matched” means in this context.
It is common (and conceptually convenient) to think of the microscopic arrange-
ment of alveoli and pulmonary capillaries as one in which alveoli all reside at the end
of conducting airways and are more or less equivalent from a mass transport view-
point (Fig. 7.15A, color plate). Unfortunately, this “parallel ventilation/parallel
A B
100
EXTRACT OF LUNG
EXTRACT OF LUNG FROM NEWBORN INFANT
FROM INFANT WITH DYING FROM NON-
RESPIRATORY DISTRESS PULMONARY CAUSES
% Surface area
60
20
0 10 20 30 40 0 10 20 30 40
Surface tension (dynes/cm) Surface tension (dynes/cm)
Figure 7.14
Surface tension measurements made on extracts from infant lungs using the apparatus shown in Fig. 7.11. Comparison
of the measurements in lung from an infant suffering from respiratory distress syndrome (A) with those from an infant
without pulmonary distress (B), and with the data shown in Fig. 7.12, shows that the minimum attainable surface tension
is elevated in the lung extract from the child with respiratory distress. From Clements [9] with permission of the
American Physiological Society.
perfusion” model is incorrect; in mammals, the alveoli within an acinus reside

in the last six to nine generations of airways, so that oxygen is continually being
extracted from acinar air as one moves deeper into the lung (Fig. 7.15B, color plate)
[11]. This means that it is more appropriate to think of the acinus as a “series ven-
tilation/parallel perfusion” system, a distinction that has important consequences
for air-side mass transfer, as we will see. Let us begin our analysis of this system
by considering blood-side mass transfer.
7.4.1 Blood-side acinar mass transfer
Based on the discussion above, we will treat all alveoli as being essentially equiv-
alent from the viewpoint of blood-side mass transfer, with the exception that we
will allow the driving potential (alveolar O2 concentration) to depend on the airway
generation. Some other salient features of mass transfer in a single alveolus are:
r blood flows in a sheet-like network of capillaries [12] that are in direct contact
with the alveolar wall on both sides
r the thin tissue layer intervening between blood and air allows gases (e.g., O2 and
CO2 ) to cross, but under normal circumstances, prevents blood from entering
the alveolus; it can therefore be thought of as a semipermeable membrane.
Semi-permeable wall Blood velocity U
Gas flux Blood in capillary lumen
Air in alveolus
y = Ro
y
x
y = −Ro
x=0 x=L
Air in alveolus
Effective blood gas

concentration c Alveolar gas
concentration
cout
c in
x
Figure 7.16
Schematic of mass transfer to an alveolar capillary (not to scale). Blood flows from left to right in the capillary, here
shown as a central channel, where Ro is the radius of a single pulmonary capillary. Alveolar air contacts the upper and
lower boundaries of the channel, which are assumed to be gas permeable. The graph shows how blood gas concentration
changes with axial position in the capillary, in this case approaching but not quite reaching the alveolar gas concentration.
Based on these features, we model the capillaries in contact with the alveoli as a
two-dimensional channel with semipermeable walls. The height of the channel is
2Ro , where Ro is the radius of a single pulmonary capillary. Alveolar air is in direct
contact with the upper and lower semipermeable walls, and blood flows through
the channel with mean velocity U (Fig. 7.16).
We consider the transport of a gas from the alveolar air to the blood. Obviously
this is relevant to the transport of O2 , but we will see shortly that the situation for
O2 is remarkably complex, essentially because the binding of O2 to hemoglobin

depends in a non-linear manner on O2 partial pressure and a number of other
factors. Therefore, to get started, let us simply consider a generic gas whose total
solubility in blood is proportional to its partial pressure. We will denote the local
effective concentration of this gas in the blood by c(x, y, t); it has dimensions of
moles of gas per volume of blood. This gas is able to cross the semipermeable
capillary walls by diffusion. Its mean concentration in the incoming blood stream
is cin , and in the blood leaving the capillary sheet it has mean concentration cout .
We wish to solve for the blood gas concentration, c(x, y, t). A complete analysis
of this problem is rather difficult, since c is in general a function of position
(x and y) and time t. We will therefore simplify the problem by making several
assumptions.
1. We assume that the blood is well mixed in the cross-stream (y) direction. This
implies that c is not a function of y. This is actually a poor approximation,
since it ignores the mass transfer boundary layer that forms along the semiper-
meable walls of the pulmonary capillaries. We will discuss this assumption on
p. 304.
2. We assume steady-state behavior, so that c is not a function of time. This means
that the pulsatile blood flow is replaced by a steady mean flow, and that the
time-varying concentration of gases in the alveolar air is replaced by a steady
mean concentration. This will also be discussed in more detail on p. 305.
We are left with c depending only on axial position, x. This dependence on x is

central to the physics of the problem, since gas is continually leaving the alveolus,
diffusing across the semipermeable membrane, and being taken up by the blood.
The more rapidly the gas can diffuse across the membrane, the more quickly c will
increase with x. An important parameter in this analysis is therefore expected to
be the flux of gas across the walls, J, defined by
moles gas diffusing across channel boundaries per unit time
J= . (7.2)
unit wall area
The driving force for the diffusional gas flux J is the gas concentration difference
between the alveolar air and the blood. Since c depends on x, this difference also
depends on x, and we therefore expect J to depend on x. Consequently, we must
formulate two equations to solve for the two unknowns, c(x) and J (x).
The first step is to perform a mass balance on the gas (not the blood) using a
control volume of length δx, as shown in Fig. 7.17. Here we exploit the symmetry
of the problem and consider a control volume in the upper half of the channel
only. The volume flow rate of blood into this control volume (per unit depth into
the page) is Q blood = U Ro . Since there is no gas transported across the center line
J (x )
Blood velocity U Blood velocity U

Gas concentration c Ro Gas concentration c + dc
Centerline
Control volume
δx
Figure 7.17
Control volume for analyzing alveolar gas transfer. Blood enters from the left at mean velocity U. Gas concentration in the
incoming blood is c; in the outgoing blood it is c + δc. The control volume is δx long and one unit deep (into the
page). R o is the radius of a single pulmonary capillary.
(symmetry!), a mass balance per unit depth into the page yields
c(x)Q blood + J (x)δx = (c(x) + δc) Q blood . (7.3)
The various terms in Equation (7.3) represent, from left to right, the mass flow
rate of gas entering the left-hand edge of the control volume; the mass flow rate
of gas entering the top of the control volume through the semipermeable upper
channel wall; and the mass flow rate of gas leaving the right-hand edge of the
control volume. Rearranging Equation (7.3) and taking the limit as δx → 0 yields
dc dc
J = Q blood = Ro U . (7.4)
dx dx
The second step is to relate J (x) to c(x) by using Fick’s law, which states that
the diffusional flux, J, of a quantity is proportional to the concentration gradient
of that quantity in the direction of diffusion. The proportionality coefficient is the
diffusivity, D. In the present case, the concentration gradient is the concentration
difference across the upper wall of the channel divided by the thickness of the
upper wall. Taking account of the fact that the flux is in the direction from higher
concentration to lower, we can therefore write the diffusional flux as
calv − c(x)
J (x) = D (7.5)
y
where calv is the concentration of gas in the alveolar air adjacent to the upper
channel wall and y is the thickness of the upper channel wall.
What is the meaning of the term calv ? Put another way, the concentration c
is defined in terms of moles of gas per volume of blood, so how can we talk
about c in the alveolar air, where there is no blood? To resolve this question, we
need to take a detour and discuss the concept of equilibrium between two phases.
Suppose we have a gas mixture containing a gas X and that this gas mixture is
allowed to come to equilibrium with a liquid in which the gas X is soluble. In the
simplest case, the equilibrium concentration of gas X in solution in the liquid, cX ,
will be linearly proportional to the partial pressure3 of gas X in the mixture, pX ,
i.e.,
cX = βX pX (7.6)
where the constant βX is known as the solubility coefficient for gas X.

For example, the solubility of O2 in water at 37 ◦ C is approximately
1.41 × 10−9 mol O2 /(cm3 mmHg) [2].4 If Equation (7.6) – which is nothing other
than Henry’s law written in a different way – holds, then it is clear that there is a
linear relationship between the concentration of gas X dissolved in a liquid and the
mole fraction of gas X in the equilibrated gas phase. Hence, if all references to the
gas phase are understood to be at a constant reference pressure and temperature,
cX can be thought of as a surrogate measure of the concentration of X in the gas
phase. It is in this sense that we use calv in Equation (7.5). By the same logic, the
reader will appreciate that we can use the partial pressure, pX , to represent the
concentration of gas X in solution in the liquid. It is common to adopt the second
approach, and we will do so here. Specifically, we will now solve for the partial
gas pressure p(x) in the blood, with alveolar partial pressure palv , etc.
Let us now return to Equation (7.5). In general, palv is a function of position
and time. However, we will assume that the alveolus is large and its contents are
well mixed. Although this assumption is not particularly accurate, it permits us to
approximate palv as a constant; it will be further discussed below.
The third step is to combine Equations (7.4), (7.5), and (7.6) to obtain
the following first-order linear differential equation with constant coefficients
for p(x):
dp D
Ro U = [ palv − p(x)] . (7.7)
dx y
3
The reader will recall that in a mixture of ideal gases, the partial pressure of gas X is equal to the mole fraction of gas X
multiplied by the total pressure of the gas mixture.
4
Sometimes solubility coefficients are given in terms of volume of gas rather than moles of gas, in which case the
coefficient is known as the Bunsen solubility coefficient. But one must be careful when using these numbers, since the
volume of gas depends on the temperature and pressure, and it should therefore be clearly stated what reference state is
being used. Even worse, the solubility itself depends on temperature, so there are two temperatures involved. It is
common practice in the literature to report Bunsen solubility coefficients at 37 ◦ C with the gas volume referred back to
standard temperature and pressure (i.e., 0 ◦ C and 1 atm). It is also interesting that the solubility of gases in blood plasma
is slightly less than in pure water. For O2 at 37 ◦ C, the plasma solubility is about 90% of that for water [13]; for CO2 the
serum:water solubility ratio is about 93% [14,15].
Since Equation (7.7) is first order, it requires one boundary condition, which is that
the inlet blood gas partial pressure must equal pin , i.e.,
p(x = 0) = pin . (7.8)
The fourth step is to solve Equation (7.7). To accomplish this it is convenient to

define a dimensionless partial pressure p̂ by
p − pin
p̂ = . (7.9)
palv − pin
Since p(x) can never be greater than palv , p̂ must always be 1. Similarly, the
minimum possible value of p(x) is pin , so that p̂ 0. Finally, p̂ is linearly propor-
tional to p. Therefore, p̂ is a convenient non-dimensional measure of concentration
lying in the range 0 p̂ 1 such that:
r p̂ close to 1 means that blood is nearly fully loaded with gas

r p̂ close to 0 means that blood is carrying the minimum amount of gas.
We substitute p̂ into Equation (7.7), and after some algebraic manipulation obtain

d p̂ D
+ ( p̂ − 1) = 0. (7.10)
dx Ro U y
The boundary condition is p̂ = 0 at x = 0. This system has solution
p̂ = 1 − e−x/L char (7.11)
where L char is a constant with dimensions of length equal to
U Ro y
L char = . (7.12)
D
L char is the e-folding length for mass transfer: that is, it is the length over which
the concentration c will change by a factor of e = 2.718 . . . .
Equation (7.11) indicates that the blood gas partial pressure increases exponen-
tially along the capillary. One property of the exponential function is that it goes
to 0 very quickly for negative arguments, so that p̂ approaches 1 quickly (Table
7.1). In the context of the lung this implies that the gas tension in the pulmonary
capillaries rapidly approaches palv , to within 1% for x/L char > 5. Therefore, from
Table 7.1. Normalized blood gas partial

pressure, p̂, in the alveolar model as a function
of dimensionless axial position, x/Lchar
x/L char p̂
1 0.63
2 0.86
3 0.95
5 0.993
10 0.99996
the mass transfer viewpoint, the optimal capillary length is approximately 4L char . If
the capillary is shorter, the blood cannot take up the maximum amount of gas, while
a longer capillary gives no mass transfer benefit but has increased flow resistance.
In the above analysis, we have considered a gas that is being transferred from
the alveolar air to the blood. Luckily, the same analysis works for the case of a gas
going in the opposite direction, such as CO2 (although see below for complexities
of CO2 transport). In that case, the definition of p̂ is altered, but the variation of
concentration with length takes the form of a decaying exponential with exactly
the same characteristic length as was derived above.
It is interesting to see what optimal capillary length is predicted by this analysis.
We use the following numbers, suitable for humans at rest:
r Capillary radius Ro has been morphometrically estimated to be about 4 µm =
4 × 10−4 cm [3,16].
r The blood speed, U, is not easy to estimate in humans. In dogs, we can estimate
that the speed is approximately 0.1 cm/s [17,18], and we will assume that a
similar value holds in humans. This is of the same order as blood speed in the
systemic capillaries [19].
r The diffusivity, D, should include hindered diffusion in the alveolar wall, so
diffusion across the alveolar wall will be slightly slower than that in water. For
O2 and CO2 , the diffusion coefficients in water at 25 ◦ C are 2.10 × 10−5 cm2 /s
and 1.92 × 10−5 cm2 /s, respectively. Weibel [2] gave a value of 1 × 10−5 cm2 /s
for O2 in connective tissue at 37 ◦ C, and we will take this as being typical of
the gas species of interest.
r The effective tissue thickness must account for local variations in thickness and
the fact that gas will be transported preferentially across thinner wall regions.
Weibel [2] discussed this in detail and indicated that the appropriate value is
y ≈ 0.6 µm (i.e., 6 × 10−5 cm).
These numbers give L char = 2.4 × 10−4 cm, or 2.4 µm. The actual capillary
length L is of order several hundred microns [18] to 500 µm [20], which is many
times L char . This analysis therefore suggests that there is significant “extra mass
transfer capacity” in the pulmonary capillaries. However, we should be a little
careful about this conclusion, for the following reasons:
r There is a large amount of uncertainty in some of the above parameters, arising
from several sources. First, what does “capillary length” mean in a capillary
network like the pulmonary capillaries, where there are extensive interconnec-
tions between capillary segments? We have taken it to be the distance from the
terminal pulmonary arteriole to the pulmonary venule, but even this is some-
times not easy to define in an unambiguous manner. Second, even if we can
define the relevant quantities, it is very difficult to measure them at the single
capillary level, especially in humans; see [3,16,20,21] for an appreciation of
some of these difficulties.
r The values taken above are averages, but there is a great deal of heterogeneity
in the lung. Because blood flow and gas transport will tend to follow a path
of least resistance, incorporation of such heterogeneity into models is a major
challenge [22].
r We have neglected mass transfer resistances in the plasma and within red cells
and have assumed that the kinetics of gas binding in the red cell are very fast.
This turns out to be not quite right, as will be discussed in more detail below.
The net effect of incorporating these effects would be to significantly increase
L char .
Putting this all together, we can say that the value for L char derived above almost
certainly underestimates the true value. Nonetheless, even if we were wrong by
a factor of 10, our analysis still predicts that there is “extra capacity.” But this is
not a bad thing! Remember that the calculation above was for the resting state.
During exercise, the mass transfer requirements are more severe, and it is therefore
appropriate to have extra capacity for such times.5
Complexities associated with O2 and CO2 transport

A key assumption in the above analysis was that the partial pressure of the gas was
directly proportional to the blood concentration of the gas. However, for the gas
that we care most about – O2 – this is quite a poor assumption. Even the situation
with CO2 is more complex than we have considered above.
5
Actually, the situation during exercise is somewhat more complicated, since under-perfused pulmonary capillaries are
“recruited” during exercise. Consequently, there is not a one-to-one relationship between cardiac output and U during
exercise. However, the comments made above still hold in a qualitative fashion.
We will first consider the situation with O2 . Here the difficulty arises because
O2 in the blood is carried both by the aqueous phase (plasma) and by hemoglobin
within the red cells. In fact, most O2 is carried bound to hemoglobin, so Equation
(7.6) must be modified to read
cO2 = βO2 pO2 + 4cHb S( pO2 ) pO2 (7.13)
where the second term represents the O2 bound to hemoglobin. The amount of
O2 carried in this manner depends on the hemoglobin concentration, cHb , and
the oxygen saturation of hemoglobin, S( pO2 ). The factor of 4 arises because one
molecule of hemoglobin can bind four molecules of oxygen. The oxygen saturation
function equals the fraction of hemoglobin that is bound to oxygen, and is a pure
number between 0 and 1. It can be approximated by the Hill equation
n
pO2-
p50
S( pO2 ) = - n (7.14)
1 + pO2 p50
where empirically n = 2.7 and p50 , the partial pressure at which hemoglobin is
50% saturated by O2 , is approximately 27.2 mmHg. Equation (7.14) shows that O2
blood concentration depends on O2 partial pressure in a highly non-linear manner
(Fig. 7.18), so it is no longer straightforward to relate gas-phase partial pressures
to blood-phase gas concentrations. Additionally, O2 saturation depends on blood
pH and CO2 concentration (the Bohr effect [2]), both of which vary with position
in the pulmonary capillary. The net effect is that O2 binds hemoglobin much more
avidly near the entrance to the pulmonary capillaries than would be expected by
linear theory, which is helpful in that it facilitates rapid uptake of O2 by blood in
the lungs.
Because of these complications, there is a non-exponential relation between p
and x for O2 (Fig. 7.19A). However, the scaling in the above analysis (i.e., the
dependence of L char as given by Equation [7.12]) is still approximately valid. For
a more complete discussion of O2 mass transfer, the reader is referred to the books
by Weibel [2] and Cooney [15].
The situation for CO2 transport is slightly simpler. CO2 is about 24 times more
soluble in the aqueous phase than is O2 , and it is carried in the blood in three main
forms: as a dissolved gas in the plasma (7% of the total), bound to hemoglobin
(30%), and in the form of bicarbonate ions in the red cells (63%) [15]. The net effect
is that there is less non-linearity in the relationship between CO2 partial pressure
and CO2 concentration in blood than there is for O2 , and hence the partial pressure
distribution more closely approximates an exponential for CO2 (Fig. 7.19B).
Total O2 22
100
Oxygen content (ml O2 per 100 ml blood)

Oxygen saturation of hemoglobin, S( pO ), %
18
2
80 O2 combined with
hemoglobin
14
60
10
40
6
20
Dissolved O2 2
0 20 40 60 80 100 600
Partial O2 pressure, pO (mm Hg)
2
Figure 7.18
Oxygen-binding characteristics of blood. The lower dashed line represents the O2 -carrying capacity of plasma (right
scale); it can be seen that at a typical alveolar O2 partial pressure of 100 mmHg, plasma can only carry about 0.24 ml of
O2 at standard temperature and pressure (STP) per 100 ml plasma. The S-shaped solid line shows the carrying capacity
of hemoglobin (right scale) and the O2 saturation function (left scale). The upper dashed line shows the total O2 -carrying
capacity of blood (right scale), which is about 20.8 ml O2 at STP per 100 ml blood. It can be appreciated how critical the
O2 -carrying function of hemoglobin is, since without it the blood simply could not carry enough O2 to peripheral tissues
without having an enormous cardiac output. From Weibel [2], based on West, Respiratory Physiology. The Essentials,
Williams & Wilkins, 1974, with permission.
Assumptions of the blood-side model, and more sophisticated models

At this point we should revisit our assumptions. There were two main ones.
The first assumption was that the alveolar wall was the major mass transfer
barrier, so we could neglect mass transfer effects in the capillary, thereby neglecting
the dependence of c on y. There are three such capillary-side mass transfer effects
for O2 : transfer through the plasma to the red cell surface, transport within the red
cell, and the kinetics of O2 binding to hemoglobin. The latter two are often grouped
together in the literature.
Classical analysis would suggest that we should consider the mass transfer
boundary layer in the blood to estimate the resistance for O2 transport through
A B
Alveolus pO = 104 mmHg Alveolus pCO = 40 mmHg
2
2
pO = 40 mmHg Pulmonary pO2 = 104 mmHg pCO2 = 45 mmHg Pulmonary pCO2 = 40 mmHg
2
capillary capillary
Arterial Venous Arterial

end end Venous
110 Alveolar pO2 level end
45 end
100
44
Carbon dioxide
90
mmHg
mmHg
43
Bl o
80
Oxygen
od
O2
dp
pC
70
Bloo
O
2
42
60
41
50
40 40
Alveolar pCO2 level
Figure 7.19
Distributions in a pulmonary capillary of O2 (A) and CO2 (B). Note that the distribution of O2 is not an exponential, for
reasons described in the text. For CO2 , there is an exponential decay of partial pressure ( p CO2 ) with axial distance down
the capillary. In this plot, the gas concentration is expressed in terms of the gas partial pressure for component i, p i ,
given by p i = mole fraction i × total pressure. This can also be written as p i = p tot c i /( j c j ), where p tot is the total
pressure, c i is the molar concentration of gas species i, and j refers to a sum over all species present. From Guyton
[23], based on calculations reported in Milhorn and Pulley [24].
the plasma. However, this is not really appropriate, since red cells are passing
through the capillary in the near vicinity of the wall, hence disturbing this mass
transfer boundary layer. It is more appropriate to think of a simple diffusional
process across the plasma layer between the capillary wall and the red cell. One
can then think of the plasma layer as being an extension of the capillary wall, so
the effective thickness y becomes the sum of these two layers. Weibel et al. [20]
estimated the thickness of this composite layer to be approximately 1.1 µm, which
approximately doubles the value of L char obtained from the above analysis.
More serious is the issue of O2 transport and binding within the red cell. A
detailed analysis of this effect is beyond the scope of this book, but Weibel has
considered these factors in a model of pulmonary capillary mass transfer that
implicitly (but not explicitly) includes the variation of O2 tension along the length of
the capillary [2]. His numbers, while only approximate, suggest that mass transfer
effects in the red cell can actually dominate the overall mass transfer process, and
the effective value of L char would be three to five times larger than our computed
value if these effects were taken into account.
The second assumption was that the system could be described as steady, so that
neither c nor calv depended on time. The magnitude of the error introduced by this
assumption depends on the time scale for mass transport across the alveolar wall
compared with the time scale of changes in alveolar O2 and CO2 concentration. If
the mass transport is fast, then the process is quasi-steady, and the above model
will be approximately correct if all concentrations, fluxes, etc. are taken as time-
averaged quantities. The time scale for transport in our model is the time required
to diffuse across the alveolar wall. The physics of the diffusion process show that
this is y 2 /2D, which is about 10−4 s. This is very fast compared with the time
scale of changes in alveolar O2 and CO2 concentration, which is the breathing
period (about 5 s). Therefore, this assumption is acceptable. Furthermore, at least
for the case of O2 , the variation in alveolar O2 concentration is rather modest,
ranging from about 100 to 105 mmHg over the breathing cycle [2]. Considering
that the O2 tension in the blood entering the capillaries is only approximately
40 mmHg, it is acceptable to replace the alveolar O2 partial pressure by its mean
value.
We can see that there is room to develop more sophisticated models than the one
described above. For example, it turns out that the hematocrit in the pulmonary
capillaries is important; since most O2 is taken up into red cells, and red cells do
not fill the entire capillary, only part of the capillary wall is effectively “available”
for O2 transport. This can be accounted for by computing an effective capillary
wall area weighted by the presence of red cells [20] or by more complex models
where individual red cells are considered [22,25,26].
In closing, we would like to mention another way to think about mass transfer
at the capillary level. If we consider the entire lung, we expect that the transport
rate of O2 (or any other gas) from air to the blood, Q O2 , can be written as
Q O2 = D L O2 ( p AO2 − p̄CO2 ) (7.15)
where p AO2 is the O2 partial pressure in the alveolar air, p̄CO2 is a suitable mean O2
partial pressure in the pulmonary capillaries, and D LO2 is known as the pulmonary
diffusing capacity. In a 70 kg healthy adult human, experimental measurements
show that D LO2 is about 30 ml O2 /(min mmHg) at rest and 100 ml O2 /(min mmHg)
during heavy exercise [2]. Now, in principle, it should be possible to relate the
pulmonary diffusing capacity to the characteristics of the capillaries in the lung.
The calculation proceeds in a manner similar to that of a single capillary, except
that it is now the entire capillary surface area in the lung that is relevant. The
rate of uptake of O2 by hemoglobin – which we did not account for in our
analysis – can be estimated from measurements on whole blood. The computed
value of D LO2 obtained in this manner is close to, but somewhat larger than, the
measured value. For full details, the reader is referred to the wonderful book by
Weibel [2], and its update [20].
7.4.2 Air-side acinar mass transfer
Now let us shift our attention to mass transfer on the air side. Thanks to the large
diffusivity of O2 in air (approximately 0.2 cm2 /s [27]), diffusion on the length
scale of a single alveolus is relatively rapid, with a diffusion time of only about
0.01 s across a 300 µm alveolus. This would suggest that the air side should con-
tribute little mass transfer resistance to the overall air/blood mass transfer process.
However, the situation is a little more complex than this. We must recall that air
transport in the pulmonary airways occurs by a combination of convection and
diffusion: transport in the larger airways is convection dominated, while diffusion
becomes increasingly important as we progress to the smaller airways. It turns out
that diffusion takes over from convection somewhere around airway generation 17
in humans [11]. That means that O2 has to diffuse all the way from generation 17 to
the terminal alveoli. We will refer to this as a “gas exchange unit;” it corresponds
to one eighth of an acinus. With reference to Fig. 7.15B (color plate), we can see
that the relevant length scale for diffusive O2 transfer is therefore not an alveolar
diameter but rather the size of the gas exchange unit.
To understand air-side mass transfer better, let us consider the fate of an O2
molecule in an acinus. It is transported by convection to the proximal end of
the gas exchange unit (generation 17) and then diffuses deeper into the airways.
Along the way, it has some probability of coming close to the alveolar walls
and entering the blood. If the rate at which O2 is transported along the acinus is
very slow compared with the uptake rate into the blood, it can be seen that the
O2 concentration will significantly decrease as one moves distally towards the
terminal alveoli. This phenomenon is known as screening [11] and is undesirable,
since it means that distal alveoli “see” reduced O2 concentration and therefore
are not working optimally. To understand this phenomenon better, we can define
a screening length , which can be interpreted as the characteristic size of a gas
exchange unit at which screening begins to occur. The question then becomes what
the ratio of to the actual size of the gas exchange units is.
Based on the above discussion, we expect that the screening length will depend
on the diffusivity of O2 in air, DO2 , and on the ease with which O2 can cross the
acinar wall to enter the blood [11]. The latter quantity can be characterized by the
permeability of the acinar wall to O2 , WO2 , defined via
NO2 = WO2 A cO2 (7.16)
where NO2 is the rate at which O2 is transported across the wall of the gas exchange
unit (moles O2 per unit time), A is the surface area of the wall of the gas exchange
unit, and cO2 is the O2 concentration difference between alveolar air and blood.
NO2 has dimensions of length/time, and hence dimensional analysis shows that
WO2 /DO2 is a -group. Just as in Sections 4.3.4 and 5.1.2, we can then argue that
this single -group should be constant, which implies that the screening length
can be estimated by
D O2
∼ (7.17)
WO2
We can estimate WO2 from the properties of the acinar wall (thickness, diffusivity of
O2 in connective tissue, solubility of O2 in water) just as in Section 7.4.1. Sapoval
et al. [11] reported relevant parameters for several species and derived a value of
of approximately 28 cm in humans.
We need to compare this value of with the characteristic size of the gas
exchange unit to determine whether or not screening occurs. The relevant charac-
teristic size turns out not to be the flow pathway length in the gas exchange unit;
instead, we have to take account of the convoluted surface of the acinar wall, and
hence the relevant length scale is [11]
A
L∼ (7.18)
D
where D is the diameter of a circle that encloses a single gas exchange unit. In
humans, L is approximately 30 cm, which almost exactly equals . This close
agreement between and L suggests that the geometry of the acinus is such that
screening is minimal, and hence that air-side O2 transfer is optimized. A more
detailed calculation [11] showed that some screening occurs, but its effects are not
severe.
This is a very satisfying result, but what happens during exercise, when O2 needs
become greater and more O2 crosses the acinar wall per unit area? The above
analysis suggests that screening would occur in such a situation, which would be
very undesirable. However, we must recall that ventilation rate increases under
such conditions, with a corresponding increase in air velocities. The net effect is
that air is convected further into the acinus, so that the point at which diffusion
takes over is pushed deeper into the lung (Fig. 7.20). This means that the effective
length scale of the gas exchange unit L decreases to match the decreased screening
length . The interested reader is referred to the papers of Sapoval and coworkers
[11,28,29] for more details.
7.4.3 Whole lung mass transfer
We now shift attention from the alveolar level to consider gas transfer for the entire
lung. To do so, we need first to understand a few basic facts about gas volumes
transferred in and out of the lungs during normal breathing.
B C
Figure 7.20
Schematic description of different ways in which the acinar gas exchange unit might function. The arrows symbolize
convection; the end of the arrows is the transition point after which diffusional transport takes over. The dot represents O2
molecules. (A) The diffusivity of O2 is very large and essentially no screening occurs; consequently O2 concentrations in
the terminal alveoli are close to those in the large airways. (B) Here screening is occurring; significant amounts of O2 are
removed from air before it reaches the terminal alveoli. (C) In exercise, increased flow rates shift the convective/diffusive
transition point to more distal airways and screening is minimized. From Sapoval et al. [11]. Copyright 2002 National
Academy of Sciences, U.S.A.
Total lung volume is approximately 6 liters. However, only a small fraction

(500 ml at rest) of this total volume is exchanged with each breath. The volume
exchanged per breath is the tidal volume, and it is exchanged approximately 12
times per minute. In addition to the tidal volume, other relevant terms are listed
below (see also Fig. 7.21).
Residual volume. The volume of air remaining in the lungs after maximal expi-
ration, normally approximately 1200 ml.
Expiratory reserve volume. The volume of air that can be exhaled after normal
exhalation of one tidal volume, normally approximately 1200 ml.
Inspiratory reserve volume. The volume of air that can be inhaled after normal
inhalation of one tidal volume, normally approximately 3100 ml.
Figure 7.21
Lung volumes and capacities. The left part of the figure shows an idealized spirometer tracing for three normal breathing
cycles, a maximal inspiration and expiration, and two more normal breathing cycles. When the subject inspires, the
spirometer trace moves up; with expiration it moves down. See text for definition of terms on the right side of the figure.
After Vander et al. [1]. Reproduced with kind permission of The McGraw-Hill Companies.
Vital capacity. The sum of tidal volume plus inspiratory reserve volume plus
expiratory reserve volume. It is the lung’s working maximum volume; note that it
is less than the total lung volume.
With this background, we now consider overall gas transfer in a normal lung at rest.
We expect the exhaled air to contain more CO2 and less O2 than the inspired air
and wish to quantify these concentration differences. This requires that we know
the body’s rates of CO2 production and O2 consumption at rest, which are related
to overall metabolic rate:
r at rest, the body’s O2 consumption rate is 250 ml/min at standard temperature

and pressure (STP, or 1 atmosphere and 273 K)
r at rest the CO2 production rate is 200 ml/min at STP.
It is conventional to refer all volumes to body temperature and pressure (BTP),

which is taken as 1 atmosphere and 310 K. At BTP, the O2 consumption and CO2
production rates are 284 and 227 ml/min, respectively.
For purposes of this calculation, we will assume that the ambient air has the
composition shown in Table 7.2, corresponding to a dry day. In all calculations
below we will assume that air is a mixture of perfect gases. For such a mixture, the
molar fraction of component i, n i , is related to the partial pressure ( pi ) and partial
Table 7.2. Air composition in inspired and expired tidal volumes. All volumes in
the table are referenced to BTP. The assumed molar fractions are in column two,
corresponding to dry ambient air (trace gases are not shown). The corresponding
partial pressures and partial volumes (in a 500 ml inspired tidal volume) are
shown in columns three and four. The partial volumes in the expired tidal volume
are shown in column 5, calculated as described in the text.
Gas Molar fraction Partial pressure Partial volume in one:

(%) (mmHg)
Inspired tidal Expired tidal
volume (ml) volume (ml)
N2 78.62 597 393.1 393.1
O2 20.84 159 104.2 80.5
CO2 0.04 0.3 0.2 19.1
H2 O 0.50 3.7 2.5 32.5
Total 100.00 760 500 525.2
volume (Vi ) of component i by

pi Vi
ni = = (7.19)
ptot Vtot
where ptot and Vtot are the total pressure and total volume (respectively) of the gas
mixture. Note that the total pressure and total volume obey

ptot = pj (7.20)
j

Vtot = Vj (7.21)
j
where the sum is over all gases in the mixture.

Given the ambient air composition in column two of Table 7.2, it is straightfor-
ward to use Equation (7.19) to compute the partial pressures and partial volumes
in a 500 ml inspired tidal volume (columns three and four). The partial volumes in
column five are calculated as follows:
r Nitrogen is not metabolized, and so its partial volume does not change.
r Oxygen is consumed at the rate of (284 ml/min)/(12 breaths/min), or
23.7 ml/ breath. Therefore, the expired air contains 104.2 − 23.7 = 80.5 ml O2 .
r Carbon dioxide is produced at the rate of (227 ml/min)/(12 breaths/min), or
18.9 ml/breath. Therefore, the expired air contains 0.2 + 18.9 = 19.1 ml CO2 .
Table 7.3. Comparison of calculated and

measured tidal volume compositions. Calculated
compositions are from column five of Table 7.2,
measured values are from Cooney [15]. With kind
permission of Taylor & Francis.
Gas Composition (%)

Calculated Measured
N2 74.8 74.5
O2 15.4 15.7
CO2 3.6 3.6
H2 O 6.2 6.2
r The water mass balance is slightly more complex. The essential fact is that the
expired air is, to a very good approximation, fully humidified. Therefore, the
partial pressure of water vapour in expired air equals the saturation pressure of
water at 310 K, which is 47 mmHg. Since partial pressures are directly propor-
tional to partial volumes, we can then write
volume water partial pressure water 47 mmHg

= = . (7.22)
volume dry gases partial pressure dry gases (760 − 47) mmHg
In Equation (7.22), we use the fact that the total pressure in the expired tidal
volume is 760 mmHg, consistent with the fact that all values in Table 7.2 are ref-
erenced to BTP. Adding the volumes in column five of N2 , O2 , and CO2 gives a
dry gas volume of 492.7 ml. Hence, Equation (7.22) can be used to compute
47
volume water = 492.7 ml = 32.5 ml, (7.23)
760 − 47
which is entered in Table 7.2 in column five. The total expired volume (at BTP)
is then calculated to be 525.2 ml.
The accuracy of this calculation can be checked by comparing the calculated

expired air composition with measured values. Table 7.3 shows that the agreement
is very good.
It is interesting to note that Table 7.2 indicates that the expired volume exceeds
the total inspired volume.6 At first sight this seems paradoxical, but it can be
explained by the fact that mass is being added to the expired air by the lungs.
6
This volume “discrepancy” is not as large as might be expected from Table 7.2. Recall that all volumes in Table 7.2 are
at BTP. However, the actual pressure during inspiration is slightly less than atmospheric; consequently, the actual
inspired tidal volume within the lungs is slightly greater than 500 ml. Similarly, during expiration the alveolar pressure
is slightly greater than 1 atm so that the actual expired tidal volume within the lungs is slightly less than 525.2 ml.
313 7.5 Particle transport in the lung
(The reader will recall that volume fluxes do not have to balance, but mass fluxes
do.) Inspection of Table 7.2 shows that a great deal of the volume “discrepancy”
between inhaled and exhaled air results from water addition by the lungs, and in
fact the lungs are a major site of water loss. For the ambient air conditions of Table
7.2, 30.0 ml (at BTP) of water are lost per breath. Assuming that the water vapor
is a perfect gas, this corresponds to
pV
n =
RT
1 atm × 30.0 ml 10−3 liter
=
82.05 × 10−3 liter atm/(mole K) × 310 K ml
−3
= 1 .18 × 10 mol H2 O (7.24)
This corresponds to a loss of 0.0212 g water per breath. Over the course of one
day, breathing at an assumed rate of 12 breaths/min, this produces a net daily loss
of 366 g water, which is approximately 15% of the body’s total daily water loss
from all sources.
7.5 Particle transport in the lung
The air that enters the conducting airways during breathing can carry particles with
it. These particles can be useful, for example when aerosol droplets containing
inflammation-reducing steroids are inhaled by a child to treat his or her asthma.
Unfortunately, they can also be harmful, such as occurs when asbestos fibers (or
other toxic materials) are inhaled by workers. A critical question in relation to
particulate inhalation is where the particles are deposited within the airways. For
example, several systems for the delivery of inhaled insulin to diabetics are now
being tested [30]; in this application, it is important to get as many of the drug-
containing particles as possible down into the small airways so as to exploit the
large surface area available for insulin transfer into the blood.
A number of authors have considered how particles are transported within the
airways; Grotberg has provided a recent review [31]. The most critical issue is
the ability of the particle to follow the curves and bends of the airways as it is
carried into the lung by the air. We can get some insight into this by considering
a small spherical particle of diameter dp , moving at velocity up in a flow with
local velocity U. We will consider the case where the Reynolds number based on
particle diameter is 1, so that the fluid drag force on the particle is given by
Stokes’ law [32]
drag = 3π µdp (U − up ) (7.25)

where µ is fluid viscosity. In its simplest form, the equation of motion for the
particle is therefore [33]
dup
m = 3π µdp (U − up ) (7.26)
dt
where m is the mass of the particle. The reader can verify that Equation (7.26)
can be cast into the dimensionless form
du∗p
St = (U∗ − u∗p ) (7.27)
dt ∗
where velocities with asterisks have been made dimensionless with respect to
a reference fluid velocity U ; t ∗ has been made dimensionless with respect to a
reference time scale for the flow, τ ; and the Stokes number (St) is given by
m ρ p dp2
St = = (7.28)
3π µdp τ 18µτ
where ρp is the density of the particle. Equation (7.27) shows that we will obtain two
very different types of limiting behavior depending on the magnitude of the Stokes
number. If St 1, it can be seen that the particle’s velocity will not change with
time; consequently, the particle will travel in a straight line and will eventually run
into an airway wall. If, however, St 1, the particle velocity up will always be very
close to the local fluid velocity U; in this case, the particle will be able to “follow” the
flow very well and will tend not to run into walls. The Stokes number, therefore, tells
us something about the ability of a particle to follow fluid path lines within the flow.
From the definition of the Stokes number (Equation [7.28]), we can see that a
large Stokes number will result if the particle is dense or, more significantly, if it is
large. The prediction is therefore that smaller particles should make it further into
the lungs than larger ones, and that that penetration efficiency should be a non-
linear function of particle diameter. In fact, for larger particles, as many as 45%
of all particles are deposited in the passages of the nose and mouth, depending
on the inhalation conditions [31]. This is usually beneficial, since we prefer to
stop particulate contaminants from entering the lower airways, but it is undesirable
when we are trying to deliver a drug or other therapeutic substance into the lung.
The strategy in such cases is to try to make the particles as small as possible. For
particles that are 2–10 µm in diameter, which typically have Stokes numbers of
the order of 0.005 to 0.5, the particles are able to enter the main bronchi, but even
then significant numbers are trapped after several generations of airway splitting,
with preferential deposition at the airway bifurcations along the carinal ridges
[31].
Despite the filtering function provided by the mouth and nose, the large volumes
of air that move in and out of the lungs mean that significant numbers of airborne
315 7.5 Particle transport in the lung
Figure 7.22
Micrographs showing the epithelial lining of a small bronchus, consisting of three cell types: ciliated (C), goblet (G), and
basal (B). The transmission electron micrograph at left shows macrophages (M) lying on the ciliated cells, as well as
underlying fibroblasts (F) and connective tissue fibers (cf). The scanning electron micrograph at right shows the extensive
cilia (C) and a goblet cell secreting a mucous droplet. Scale markers in both panels: 5 µm. Reprinted by permission of
the publisher from The Pathway for Oxygen: Structure and Function in the Mammalian Respiratory System by E. R.
Weibel, p. 237, Cambridge, MA: Harvard University Press, Copyright C 1984 by the President and Fellows of Harvard
College.
particles enter the airways and adhere to the large area presented by the airway and
acinar walls. Even for “inert” particles, continued normal lung function requires
a mechanism to deal with these materials. What does the body do with these
particles? It has several coping strategies.
1. The conducting airway walls are coated with a thin layer of mucus, which
causes particles to stick to the walls where they impact. Mucus is constantly
being synthesized by goblet cells lining the airways and is then moved upwards
towards the throat by the action of cilia, small hair-like projections of the
epithelial cells lining the airway walls (Fig. 7.22). These cilia beat back and
forth at a regular frequency to create a “ciliary escalator” that delivers mucus
(with attached and embedded particles) to the pharynx, where it is usually
swallowed with little attention. Only when excessive amounts of mucus are
present in the airways, such as during an upper respiratory tract infection or
during respiration in a dusty environment, do we notice the mucus and have
to assist the cilia by coughing.
2. Some particles (the smallest ones) pass into the alveoli, where they adhere to
the alveolar walls. Cilia are not present at this level in the lung, so the particles
must be dealt with locally. These particles are internalized by macrophages
lying on the pulmonary epithelial surface in a process known as phagocytosis.
Depending on the composition of the particle, it is then broken down to a
greater or lesser extent in the macrophage by lytic enzymes, and the breakdown
products are stored in the cell in membrane-delimited spaces called residual
bodies.
7.6 Problems
7.1 Consider a small spherical bubble of radius R.

(a) Show that the energy required to expand this sphere by a small amount
R is 2σ V /R. Here V is the increase in volume and σ is the inter-
facial tension.
(b) Estimate the time-averaged power required to overcome alveolar surface
tension during normal breathing. Take R = 150 µm, σ = 25 dynes/cm,
and breathing rate = 12 breaths/min.
(c) Repeat this calculation for the cat, where R = 50 µm and the tidal vol-
ume is 20 ml. Compare this calculated value with a rough estimate of the
power obtained from Fig. 7.23. (Take beginning of normal inspiration to
occur at 100 ml.) Is surface tension or lung tissue elasticity the dominant
restoring force in the cat lung?
7.2 This question is concerned with the energy required to inflate the lung. Specif-
ically, we wish to know what fraction of the total inflation energy is used to
overcome alveolar surface tension forces, and what fraction is used to over-
come parenchymal elasticity.
(a) How much energy is required to inflate a spherical bubble from radius
R1 to radius R2 ? The surface tension coefficient is constant and equal
to σ . Hint: think in terms of a pressure–volume relationship.
317 7.6 Problems
Saline Air
200
150
Volume (ml)
100
50
0 4 8 12 16 20
Pressure (cm H2O)
Figure 7.23
For Problem 7.1. A pressure–volume curve for a cat’s lungs as they were inflated and then deflated, first with air and then
with saline. Arrows indicate direction of inflation/deflation. Modified from Radford [34] with permission of the American
Physiological Society.
(b) Considering surface tension effects only, how much energy is required
to inflate all 300 million alveoli in the lungs? Assume that the total
alveolar volume before inspiration is 2.5 liters, that the tidal volume
is 500 ml, that all alveoli are identically sized spheres, and that the
effective surface tension coefficient during inspiration is constant and
equal to 35 dynes/cm.
(c) Idealized data from the air inflation of a pair of lungs are plotted in
Fig. 7.24. Based on your calculations from parts (a) and (b), plot (to
scale) on the same graph the pressure–volume curve expected from
surface tension effects only. What fraction of the total energy required
to inflate the lungs is from surface tension effects?
7.3 A balloon is surrounded by a tank of liquid at negative pressure and is con-
nected to the atmosphere by a tube of length L and cross-sectional area A
(Fig. 7.25). The pressure inside the balloon p oscillates above and below
atmospheric pressure causing small changes in the balloon volume V. The
elasticity of the balloon is characterized by its compliance C, defined by
p = V /C
(a) Derive a second-order differential equation for V (t), assuming that
(i) the pressure differential along the tube accelerates the air in the tube
and is not used to overcome entrance, exit, or tube losses; and (ii) the air
density ρ is constant. From the equation, show that the natural frequency
√
of the system is A/ρ LC.
10000
Inflation pressure (dynes/cm2)
5000
0
2000 2500 3000 3500
Alveolar volume (ml)
Figure 7.24
For Problem 7.2. Note that inflation pressure equals alveolar pressure minus intrapleural pressure.
Atmosphere
Length = L; Cross-sectional area = A
Air at
pressure p
Liquid Balloon
Figure 7.25
For Problem 7.3.
(b) For a 70 kg man, A/L is approximately 0.001 m. The equivalent value

for a 12 kg dog would be approximately [12/70]1/3 of the value, or
5.6 × 10−4 m. The compliance of dog lungs is approximately
0.029 l/cmH2 O. Estimate the natural frequency of a dog’s breathing
using the formula developed in (a). Measurements indicate that dogs
with a body mass of 12 kg pant at about 5.3 Hz. Comment briefly on any
differences between your answer and the measured frequency.
319 7.6 Problems
Figure 7.26
For Problem 7.4.
7.4 A lung is inflated with water and then with air. The pressure–volume curves for
these two inflation procedures are shown in Fig. 7.26, with the right-pointing
arrow representing inflation and the left-pointing arrow representing defla-
tion. Assume that the lung has 150 × 106 identical alveoli, and that alveoli
make up a constant 85% fraction of total lung volume. Based on these curves,
graph the relationship between surface tension coefficient (in dynes/cm) and
alveolar radius (in microns). Your graph should be quantitatively correct.
This is best accomplished by choosing some key points from Fig. 7.26, trans-
forming them to suitable values on your graph, and then interpolating by
sketching.
7.5 Figure 7.12 shows that a solution containing lung extract exhibits hysteresis
in its surface tension versus area relationship. In other words, the surface
tension is higher during inflation of the lung than during deflation.
(a) By recalling that mechanical work can be expressed as ∫ pdV , show that
the work required to inflate all the alveoli in the lung against the effects
of surface tension can be written as

Work = σ dA (7.29)
where σ is the surface tension coefficient, A is the aggregate surface

area of all alveoli in the lung, and the integral is carried out from mini-
mum surface area (start of inspiration) to maximum surface area (end of
Figure 7.27
For Problem 7.5.
inspiration). To show this result, you may assume that the pressure
outside the alveoli is constant and equal to 0 (gauge). You will find
it convenient to recall that the volume and surface area of a sphere
are V = 4πR 3 /3 and S = 4πR 2 . This implies that dV = 4π R 2 dR and
dS = 8π R dR.
(b) Assume that the surface tension versus area curve for the entire lung
over one breathing cycle can be approximated by the shape in Fig. 7.27.
Using this information, determine how much energy is dissipated in
surface tension hysteresis effects during one breathing cycle.
7.6 Consider a fluid layer A surrounded on both sides by a fluid, B. The fluid
layer can be thought of as a membrane. A species S is diffusing across this
membrane, and has concentrations c1B and c2B (in the fluid B) on the two sides
of the membrane, as shown in Fig. 7.28. It often happens that the solubility of
material S inside the membrane is different than its solubility in bulk solution
B. We therefore define a partition coefficient k as
concentration of S in material A (at equilibrium)

k= . (7.30)
concentration of S in material B
Hence k < 1 means that S is less soluble in the membrane, and k > 1
means S is more soluble. Write down an expression for the flux across the
membrane in terms of c1B , c2B , DA (diffusion coefficient of S in A) and yA .
Sketch the concentration profile. What is the effective diffusion coefficient
value with partitioning, Deff ?
321 7.6 Problems
Fluid A
Fluid B Fluid B
Concentration Concentration
c 1B ∆yA c 2B
Figure 7.28
For Problem 7.6.
Gas, cg(x )
velocity U 2 Ro
Membrane
Blood, CO2
velocity U
2 Ro
cb(x )
Figure 7.29
For Problem 7.7.
7.7 A membrane oxygenator is designed to supply O2 and to remove CO2 from

blood during surgery. Blood and a gas mixture flow from left to right in two
channels separated by a semipermeable membrane (Fig. 7.29). Blood gives
up CO2 to the gas stream continuously along the channel, so that both the
blood and gas CO2 concentrations are functions of position, x. Assume that
the channel width and average velocities are the same for both blood and gas
streams.
(a) If cg (x) is the CO2 concentration in gas stream; cb (x) is the CO2 concen-
tration in blood; U is the mean velocity of both blood and gas streams;
Deff is the effective diffusion constant of CO2 in the membrane and y
is the membrane thickness; show that the CO2 concentrations satisfy:
dcb cb − cg
=− (7.31)
dx L char
where
cb + cg = constant. (7.32)
Make and state appropriate assumptions.

(b) If the CO2 concentration in the gas is zero at x = 0 and is cb0 in the blood
at x = 0, at what x/L char is cb reduced to 60% of cb0 ?
7.8 In Section 7.4.1, we modeled gas transfer between blood in a pulmonary
capillary and an alveolus, taking account of the fact that the concentration of
gas in the blood varied with position. However, we did not account for the
time-dependent nature of the mass transfer process, which is the subject of this
question. Suppose that the CO2 concentration everywhere in a capillary can
be expressed as some spatially averaged value ccap . Because blood from the
right heart is continually flowing into the capillary, ccap can be approximated
as being constant in time. Assume that at time zero, fresh air (effective CO2
concentration of 0) enters the alveolus, that at every instant the air in the
alveolus is well mixed and does not communicate with air in the terminal
bronchioles, and that all air in the alveolus is expelled and replaced by fresh
air with every breath (breathing period = T ).
(a) Qualitatively sketch a labeled graph of alveolar CO2 concentration vs.
time for the above assumptions.
(b) If the total surface area for blood/gas exchange associated with one
alveolus is A, and the volume of the alveolus is V, derive an expression
for how the CO2 concentration in the alveolus changes with time. State
assumptions. You will want to start with a mass balance in the alveolus.
7.9 A membrane oxygenator is being designed as part of a heart–lung bypass
machine. It must be able to transfer 200 ml/min of O2 into blood flowing
at 5 l/min. Assume the blood enters the oxygenator with an effective O2
concentration of 0.1 ml O2 /ml blood.
(a) With what O2 concentration should the blood leave the oxygenator? You
can solve this question easily by thinking about an overall mass balance.
(b) One design is to make the oxygenator as a “stack” containing many
“units”, as shown Fig. 7.30. Each unit consists of a channel filled with
flowing blood, an O2 -filled channel, and flat membranes separating the
channels. The membranes are 10 cm × 10 cm by 5 µm thick, and the
height of each blood-containing channel is 1 cm. The O2 -containing
channels are filled with 100% O2 , which is equivalent to a blood con-
centration of 0.204 ml O2 /ml blood. How many membrane units are
needed to supply the required oxygen? The value for Deff of O2 in the
membranes is measured as 10−6 cm2 /s.
323 7.6 Problems
Figure 7.30
For Problem 7.9.
Pressures p are constant and identical
Outlet
Inlet dialysate
cd
x Water
Outlet
Inlet
Q (x ), c (x )
Blood
Figure 7.31
For Problem 7.10.
7.10 A dialyzer is shown in Fig. 7.31. The dialysate solution has glucose added
to it so that it is hypertonic (has an osmotic pressure greater than that of
blood). This causes water to be drawn into the dialysate, even though the
pressures in the blood and dialysate are identical. We wish to determine
how much water will be removed from the blood plasma in this unit. For
purposes of this question, we will denote the volume flow rate of blood per
unit depth into the page by Q(x), the molar concentration of osmotically
active components in the blood by c(x), and the molar concentration of
osmotically active substances in the dialysate by cd . We will assume that
only water can cross the membrane. We will also assume that blood and
dialysate obey van’t Hoff’s law for osmotic pressure.
(a) Briefly state in words why c and Q change with axial location, x.
(b) Show that the blood flow rate per unit depth into the page satisfies
dQ
= L p RT [c(x) − cd ] (7.33)
dx
where L p is the membrane permeability (volume flow of water across

the membrane per unit area per unit osmotic pressure difference).
(c) Show that the product of c(x) and Q(x) is constant. Hint: consider the
mass flow rate of osmotically active components in the blood per unit
depth into the page.
(d) If the flow rate of dialysate is large, the concentration cd can be treated
as a constant. Under this approximation, show that Q(x) satisfies

Qr − Q0 Q −Q(x) L p RT c0 Q 0
ln + 0 = x (7.34)
Q r −Q(x) Qr Q 2r
where Q 0 and c0 are the values of Q(x) and c(x) at the inlet, and Q r is
a constant defined as Q 0 c0 /cd . Hint:

Q Q r −Q
dQ = −Q r ln − Q + constant. (7.35)
Q r −Q Qr
(e) If the total blood flow rate entering the dialyzer is 250 ml/min, the dia-
lyzer is 10 cm deep into the page, and the maximum allowable water
loss from the blood is 10 ml/min, what length L should the dialyzer
be? The dialysate osmotic concentration, cd , is 0.32 mol/l, the inlet
blood osmotic concentration is 0.285 mol/l, the membrane permeabil-
ity is 1 × 10−8 cm/(s Pa), and the working temperature is 310 K. The
universal gas constant R is 8.314 J/(mol K).
7.11 Consider the transport of oxygen in stagnant blood. More specifically, sup-
pose that a large container filled with deoxygenated whole blood is sud-
denly placed in an atmosphere of pure oxygen at concentration co , where
co is constant (Fig. 7.32). Because of the very non-linear nature with which
hemoglobin binds O2 , we will approximate this situation as one in which a
“front” forms and propagates: above this interface the hemoglobin is oxy-
genated and there is free O2 in solution in the plasma, while below this
interface the free O2 concentration is zero and the hemoglobin is deoxy-
genated. The front location, z f , advances with time.
(a) Use Fick’s law for diffusion of the free O2 to argue that the concentra-
tion profile above the front is linear, and that the flux of O2 , J, is:
D
J= co (7.36)
zf
where D is the diffusivity of oxygen in whole blood. For purposes of

325 7.6 Problems
Figure 7.32
For Problem 7.11.
this portion of the question, you may assume that the front is slowly
moving compared with the time scale for diffusion.
(b) Assuming that 4 moles of O2 bind rapidly and irreversibly to 1 mole
of hemoglobin at the front, and that the concentration of deoxy-
genated hemoglobin below the front is cHb , show that the front position
satisfies:

2Dco t
zf = (7.37)
4cHb + co /2
Hint: perform an unsteady mass balance on the O2 in the container.

Do not forget to take into account O2 that is bound irreversibly to
hemoglobin.
7.12 Consider mass transfer from pulmonary capillaries to a single alveolus.
Suppose that, in addition to the normal 0.6 µm thick tissue layer between
the blood and the air, scar tissue has formed that is 1µm thick. The effective
diffusivity of CO2 in this scar tissue is 0.7 × 10−6 cm2 /s. When CO2 has to
cross both the “normal” tissue and the scar tissue, the mass transfer efficiency
of the alveolus is reduced. Considering the entire alveolus, compute the net
percentage reduction in blood-to-air CO2 transfer due to the scar tissue for
a 50 µm long capillary. Note: you do not need to re-derive equations; you
should be able to make some simple modifications to equations in the text
to get what you need. Remember that when two mass transfer barriers are
in series, their mass transfer resistances add. You may use parameter values
(except for capillary length) given in Section 7.4.1.
7.13 This question is concerned with mass transfer in the whole lung.
(a) If you switch from breathing air to breathing xenon, what is the min-
imum number of breaths after which the concentration of xenon in
your lungs is 99%, by volume? Assume that no xenon diffuses out of
the lungs, that perfect mixing occurs in the lungs, and that the xenon
environment is so large that the exhaled air does not change the xenon
concentration of 100%.
(b) If you switch back to an air environment and breathe normally, after
how many breaths is the xenon concentration in your lungs reduced to
0.1%, by volume?
7.14 You decide to cure your hiccups by breathing into a paper bag. Assume
that the bag is initially filled with 500 ml air (referenced to BTP) having
the ambient air composition given in Table 7.2. Assume a breathing rate
of 12 breaths/min, a CO2 production rate of 235 ml/min at BTP, and an
O2 consumption rate of 284 ml/min at BTP. You can also assume that you
begin by inhaling all the air in the bag (i.e., tidal volume is 500 ml for the
first breath), and that on each subsequent breath you increase your tidal
volume so as to completely empty the bag on inhalation. Compute the CO2
concentration (as a percentage) in the bag after exhalation on the 10th breath.
Although not particularly realistic, you can assume that the normal amount
of CO2 is transferred from the lungs to the expelled air with every breath,
even though the CO2 concentration in the bag is continually increasing.
7.15 You are standing on the top of Mount Everest (elevation 29 028 ft; 8708 m)
where atmospheric pressure is 235 mmHg and the ambient temperature
is 0 ◦ C. The air composition is 78.6% N2 , 20.8% O2 , 0.04% CO2 , and
0.50% H2 O. Treat the air as an ideal gas.
(a) If you remove your oxygen set, how many times per minute must you
breathe so as to satisfy your O2 requirements of 284 ml/min at BTP?
(Recall that BTP is 1 atmosphere and 37 ◦ C). Assume that you take
in tidal volumes of 1000 ml (at ambient conditions) and that you can
transfer only 30% of the O2 into your blood.
(b) Assuming that you could breathe that fast, what would the composition
of the expired air be? Your CO2 production rate is 227 ml/min at BTP,
and expired air is fully humidified. (Partial pressure of water vapour
at 37 ◦ C and 100% relative humidity is 47 mmHg.)
7.16 On a particular day, the air has the composition shown in Table 7.4. You may
assume a 530 ml tidal volume (at BTP), a breathing rate of 12 breaths/min,
an O2 consumption rate of 295 ml/min at BTP, and a CO2 production rate
of 235 ml/min at BTP.
327 7.6 Problems
Table 7.4. For Problem 7.16
Gas Partial pressure (mmHg)
N2 594
O2 156
CO2 0.3
H2 O 9.7
Total 760
Gas Molar fraction in ambient air (%)

N2 75.85
O2 20.11
CO2 0.04
H2 O 4.00
Total 100.00
The 530 ml tidal volume can be broken down into two parts: 150 ml of
dead space air, and 380 ml of alveolar air. The dead space refers to the portion
of the conducting airways where no blood/gas exchange takes place, such
as the mouth, trachea, etc. Assume that no CO2 or O2 exchange takes place
in the dead space, but that the expired dead space air is at 100% relative
humidity (at BTP). What is the composition of the expired alveolar air, and
of the expired dead space air, expressed as percentages by volume of N2 ,
CO2 , O2 , and H2 O?
7.17 During an experiment, a subject breathes in normally from the atmosphere
but breathes out into a special device that collects the water in each exhaled
breath. At the end of the experiment, the collected mass of water is 1.299 g.
During the experiment, the ambient air had the composition shown in Table
7.5. How many breaths did the subject take during the experiment? Make
and state necessary assumptions. You do not need to re-derive any formulae
appearing in the text.
7.18 While a subject was exercising vigorously the measurements in Table 7.6
were made. All measurements are referenced to BTP. Note that tidal volume
was 1000 ml (not 500 ml) and the breathing rate was 25 breaths/min. What
were the subject’s O2 consumption rate and CO2 production rate (ml gas at
STP per minute)?
Gas Molar Partial pressure Partial volume in one:

fraction (%) (mmHg)
Inspired tidal Expired tidal
volume (ml) volume (ml)
N2 78.62 597 786.2 786.2
O2 20.84 159 208.4 158.4
CO2 0.04 0.3 0.4 37.3
H2 O 0.50 3.7 5.0 64.7
Total 100.00 760 1000 1046.6
Gas Molar fraction (%)

N2 78.62
O2 20.84
CO2 0.04
H2 O 0.50
Total 100.00
7.19 An adult mouse has a respiratory rate of ∼163 breaths/min and a tidal volume
of ∼0.15 ml at BTP (37 ◦ C) [36]. The average O2 consumption for one strain
of male adult mice is 45.5 ml O2 /hour at STP [25]. For the inspired air
composition shown in Table 7.7, compute the expired air composition. You
may assume that the ratio of CO2 production/O2 consumption is the same
for humans and mice.
7.20 In respiratory analysis, it is simplest to assume that both lungs are identical
for purposes of gas exchange. This effectively implies that the air in the two
lungs can be pooled together for purposes of quantitative analysis. However,
in some cases, the volume of air entering each lung and the blood perfusion
to each lung is different. This means that the amount of CO2 added to each
lung (per breath) will be different. For purposes of this question, we will
assume that the number of moles of CO2 added to a lung is proportional to
the blood perfusion to that lung.
(a) Suppose that for a given subject the tidal volume and CO2 concentration
in expired air are measured to be 540 ml and 3.5%, respectively, and
the dead space volume is known to be 160 ml. (See Problem 7.16 for
329 References
definition of dead space volume.) If it is known that the blood perfusion

to the right lung is twice that to the left lung, while the tidal volume
entering the right lung is 1.6 times that of the left, compute the alveolar
CO2 concentrations in each lung.
(b) What criterion ensures that the alveolar CO2 concentration is equal for
both lungs?
References
1. A. J. Vander, J. H. Sherman and D. S. Luciano. Human Physiology: The Mechanisms

of Body Function, 4th edn (New York: McGraw-Hill, 1985).
2. E. R. Weibel. The Pathway for Oxygen: Structure and Function in the Mammalian
Respiratory System (Cambridge, MA: Harvard University Press, 1984).
3. E. R. Weibel. Morphometry of the Human Lung (New York: Academic Press, 1963).
4. D. W. Fawcett. Bloom and Fawcett: A Textbook of Histology (Philadephia, PA: W.B.
Saunders, 1986).
5. D. E. Schraufnagel (ed.) Electron Microscopy of the Lung (New York: Marcel
Dekker, 1990), p. 298.
6. E. R. Weibel and D. M. Gomez. Architecture of the human lung. Use of quantitative
methods establishes fundamental relations between size and number of lung
structures. Science, 137 (1962), 577–585.
7. M. Ochs, J. R. Nyengaard, A. Jung, L. Knudsen, M. Voigt et al. The number of
alveoli in the human lung. American Journal of Respiratory and Critical Care
Medicine, 169 (2004), 120–124.
8. D. L. Vawter and J. D. Humphrey. Elasticity of the lung. In Handbook of
Bioengineering, ed. R. Skalak and S. Chien. (New York: McGraw-Hill, 1987),
pp. 24.1–24.20.
9. J. A. Clements. Surface phenomena in relation to pulmonary function. Physiologist, 5
(1962), 11–28.
10. H. Bachofen, J. Hildebrandt and M. Bachofen. Pressure–volume curves of air- and
liquid-filled excised lungs: surface tension in situ. Journal of Applied Physiology, 29
(1970), 422–431.
11. B. Sapoval, M. Filoche and E. R. Weibel. Smaller is better – but not too small: a
physical scale for the design of the mammalian pulmonary acinus. Proceedings of the
National Academy of Sciences USA, 99 (2002), 10411–10416.
12. S. S. Sobin, Y. C. Fung, H. M. Tremer and T. H. Rosenquist. Elasticity of the
pulmonary alveolar microvascular sheet in the cat. Circulation Research, 30 (1972),
440–450.
13. C. Christoforides, L. H. Laasberg and J. Hedley-Whyte. Effect of temperature on
solubility of O2 in human plasma. Journal of Applied Physiology, 26 (1969),
56–60.
14. W. H. Austin, E. Lacombe, P. W. Rand and M. Chatterjee. Solubility of carbon

dioxide in serum from 15 to 38 ◦ C. Journal of Applied Physiology, 18 (1963),
301–304.
15. D. O. Cooney. Biomedical Engineering Principles (New York: Marcel Dekker,
1976).
16. B. M. Wiebe and H. Laursen. Human lung volume, alveolar surface area, and
capillary length. Microscopy Research and Technique, 32 (1995), 255–262.
17. R. L. Capen, L. P. Latham and W. W. Wagner, Jr. Comparison of direct and indirect
measurements of pulmonary capillary transit times. Journal of Applied Physiology,
62 (1987), 1150–1154.
18. N. C. Staub and E. L. Schultz. Pulmonary capillary length in dogs, cat and rabbit.
Respiration Physiology, 5 (1968), 371–378.
19. C. G. Caro, T. J. Pedley, R. C. Schroter and W. A. Seed. The Mechanics of the
Circulation (Oxford: Oxford University Press, 1978).
20. E. R. Weibel, W. J. Federspiel, F. Fryder-Doffey, C. C. Hsia, M. Konig et al.
Morphometric model for pulmonary diffusing capacity. I. Membrane diffusing
capacity. Respiration Physiology, 93 (1993), 125–149.
21. W. Huang, R. T. Yen, M. McLaurine and G. Bledsoe. Morphometry of the human
pulmonary vasculature. Journal of Applied Physiology, 81 (1996), 2123–2133.
22. A. S. Popel. A finite-element model of oxygen diffusion in the pulmonary capillaries.
Journal of Applied Physiology, 82 (1997), 1717–1718.
23. A. C. Guyton. Textbook of Medical Physiology, 4th edn (Philadelphia, PA: W. B.
Saunders, 1971).
24. H. T. Milhorn, Jr. and P. E. Pulley, Jr. A theoretical study of pulmonary capillary gas
exchange and venous admixture. Biophysical Journal, 8 (1968), 337–357.
25. A. O. Frank, C. J. Chuong and R. L. Johnson. A finite-element model of oxygen
diffusion in the pulmonary capillaries. Journal of Applied Physiology, 82 (1997),
2036–2044.
26. A. A. Merrikh and J. L. Lage. Effect of blood flow on gas transport in a pulmonary
capillary. Journal of Biomechanical Engineering, 127 (2005), 432–439.
27. E. L. Cussler. Diffusion: Mass Transfer in Fluid Systems, 2nd edn (New York:
Cambridge University Press, 1997).
28. M. Felici, M. Filoche and B. Sapoval. Diffusional screening in the human pulmonary
acinus. Journal of Applied Physiology, 94 (2003), 2010–2016.
29. E. R. Weibel, B. Sapoval and M. Filoche. Design of peripheral airways for efficient
gas exchange. Respiratory Physiology and Neurobiology, 148 (2005), 3–21.
30. D. R. Owens, B. Zinman and G. Bolli. Alternative routes of insulin delivery. Diabetic
Medicine, 20 (2003), 886–898.
31. J. B. Grotberg. Respiratory fluid mechanics and transport processes. Annual Review
of Biomedical Engineering, 3 (2001), 421–457.
32. F. M. White. Viscous Fluid Flow, 2nd edn (New York: McGraw-Hill, 1991).
33. L. A. Spielman. Particle capture from low-speed laminar flows. Annual Review of
Fluid Mechanics, 9 (1977), 297–319.
331 References
34. E. P. Radford, Jr. Recent studies of mechanical properties of mammalian lungs. In

Tissue Elasticity, ed. J. W. Remington. (Washington, DC: American Physiological
Society, 1957), pp. 177–190.
35. E. L. Green (ed.) for the Jackson Laboratory. Biology of the Laboratory Mouse,
2nd edn (New York: Dover, 1966).
36. M. R. Dohm, J. P. Hayes and T. Garland, Jr. The quantitative genetics of maximal and
basal rates of oxygen consumption in mice. Genetics, 159 (2001), 267–277.

CH 7 Resp

Uploaded by

Copyright:

Available Formats

You might also like

CH 7 Resp

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CH 7 Resp

Uploaded by

Copyright:

Available Formats

7 The respiratory system

7.1 Gross anatomy

7.1.1 The conducting airways and pulmonary vasculature

The conducting airways form a fantastically complex branching tree designed to

1984 by the President and Fellows of Harvard College.

z Name D (mm) A (cm2) Number

1 Mainstem bronchi 12.2 2.33 2

2 Lobar bronchi 8.3 2.13 4

3 Segmental bronchi 5.6 2.00 8

17 Respiratory bronchiole 1 0.40 300.0 131,072

23 Alveoli 0.25 7530.0 8,388,608

permission of Routledge/Taylor & Francis Group, LLC.

7.1.2 Associated structures

As well as the conducting airways, the pulmonary system includes a number of

7.2 Biomechanics of breathing

Inspiration. In quiet inspiration, the diaphragm contracts and moves downward,

7.3 Lung elasticity and surface tension effects

Undeformed state: low strain and low incremental Young’s modulus

Deformed state: large strain and large incremental Young’s modulus

Liquid film, pressure po

Alveolar wall (elastic)

Intrapleural pressure ppleural

Box 7.1 Analogy of an air bubble in liquid

Using Laplace’s law it is trivial to calculate the pressure difference pi − po

The relative contributions of lung parenchymal elasticity and surface tension

Clinically, surfactant is very important. It makes breathing much easier (less

7.4 Mass transfer

bind to hemoglobin in red cells. In thinking about these processes, it is convenient

perfusion” model is incorrect; in mammals, the alveoli within an acinus reside

7.4.1 Blood-side acinar mass transfer

Semi-permeable wall Blood velocity U

Gas flux Blood in capillary lumen

Effective blood gas

O2 is remarkably complex, essentially because the binding of O2 to hemoglobin

We are left with c depending only on axial position, x. This dependence on x is

Blood velocity U Blood velocity U

c(x)Q blood + J (x)δx = (c(x) + δc) Q blood . (7.3)

where the constant βX is known as the solubility coefﬁcient for gas X.

p(x = 0) = pin . (7.8)

The fourth step is to solve Equation (7.7). To accomplish this it is convenient to

r p̂ close to 1 means that blood is nearly fully loaded with gas

The boundary condition is p̂ = 0 at x = 0. This system has solution

p̂ = 1 − e−x/L char (7.11)

where L char is a constant with dimensions of length equal to

Table 7.1. Normalized blood gas partial

Complexities associated with O2 and CO2 transport

cO2 = βO2 pO2 + 4cHb S( pO2 ) pO2 (7.13)

Oxygen content (ml O2 per 100 ml blood)

Assumptions of the blood-side model, and more sophisticated models

Arterial Venous Arterial

Q O2 = D L O2 ( p AO2 − p̄CO2 ) (7.15)

7.4.2 Air-side acinar mass transfer

NO2 = WO2 A cO2 (7.16)

7.4.3 Whole lung mass transfer

Total lung volume is approximately 6 liters. However, only a small fraction

r at rest, the body’s O2 consumption rate is 250 ml/min at standard temperature

It is conventional to refer all volumes to body temperature and pressure (BTP),

Gas Molar fraction Partial pressure Partial volume in one:

NO2 = WO2 A cO2 (7.16)